The realm of high-performance computing involving supercomputers may open up more pathways for society to explore and unlock new information at a faster rate.

Countries in Asia, including China, are planning to increase their supercomputing capacities as their reliance on big data and artificial intelligence, which require more computing power, grows.

Japan, for instance, has announced plans to use its Fugaku supercomputer – the second fastest in the world – to develop a large language model (LLM) for generative AI using the Japanese language.

The tool will be released to businesses and research institutions next year, with the goal of producing more accurate results in Japanese.

According to Tecsun Yeep, managing director and principal consultant at Robust HPC, high-performance computing (HPC) uses supercomputers or clusters of supercomputers to solve complex problems at higher speeds and greater performance.

“HPC is differentiated by its ability to process data, calculate, and generate solutions at significantly faster rates and at a scale that a regular computer simply cannot handle,” he says.

Yeep explained how supercomputers are superior by using cars as an example: “One car may only transport a few passengers at a time, but having multiple cars on the road would help to transport even more passengers.

“HPC does exactly that – instead of one computer trying to solve work, you have a network of connected supercomputers in the execution.”

He claims that the use of HPC is thriving in the research field in Malaysia.

Robust HPC recently inked an MOU with the Research Centre for Crystalline Materials at Sunway University’s School of Medical and Life Sciences.

The centre, which focuses on the development of new drugs for health and well-being, is hoping to advance its research output and productivity capabilities.

He says leveraging the parallel processing capabilities of the latest graphics processing units (GPUs) will allow the institution to handle complex mathematical operations for quantum chemistry calculations.

“The use of GPUs, which can perform many calculations simultaneously, enables the calculations to be divided into smaller, parallel operations.

Number crunchers

The Frontier supercomputer, hosted at the Oak Ridge Leadership Computing Facility in Tennessee, United States, is the world’s fastest.

It is the only system with a computer power performance exceeding one exaflop per second, or a quintillion – that’s a billion billion or one followed by 18 zeros.

According to a report by IEEE Spectrum, the Frontier is being used for research in cancer, drug discovery, nuclear fusion, and more, with the goal of cutting the time required for such work from weeks to hours and from hours to seconds.

According to Yeep, HPC has the potential to address pressing issues, including climate change, public health and national security.

“Development in these areas is often tied to scientific breakthroughs and innovation, particularly in scientific computing – modelling, simulation, even artificial intelligence,” he says.

Yeep adds that HPC would enable researchers, scientists, and policymakers to tackle complex and challenging problems through its ability to actually develop, generate, and run full simulations of any scenario.

“For instance, to understand the current public health landscape, HPCs can analyse data in available medical records, clinical trials, and studies and produce deeper research into diseases and treatment options.

“In the event of a disease outbreak, HPCs could also generate models to predict the spread of the disease, which would be incredibly handy for public health officials to sufficiently prepare an action plan, on top of the potential to improve and even save lives,” he says.

He cited the DeepMind supercomputer, which helped researchers overcome complex computing challenges and eventually understand how protein molecules fold.

The discovery, dubbed AlphaFold, or protein structure prediction, is expected to help researchers look for more ways to treat diseases.

“Some organisations have taken this concept of ‘folding’, which essentially refers to running simulations of proteins, and converted it to community-led projects, which some Malaysians are also a part of.

“Volunteers contribute their PC resources to the cause, working alongside other computers in the network to carry out this computational-heavy task – in a nutshell, uniting potentially hundreds of regular computers to run simulations,” he says.

Expanding the expertise

There is a need to drive awareness for the wider use of HPC in Malaysia, says Yeep, as the increased awareness will help to increase the talent pool in the country.

“Many people are not aware of what HPCs are and how they can help transform their company’s output.

“HPC has been more associated with larger institutions or bigger, private organisations, and I believe that misconception leads to people not knowing that supercomputing has a wide scope in various fields,” he says.

About Robust HPC
Robust HPC has been a supercomputing solution provider since 2010, with clientele in Malaysia & China. Robust team is highly capable of delivering end-to-end solutions and possesses proven track records in planning and design, strategy consultation, solution implementation, user training as well as user application benchmarking/profiling/support.